Match sim with emfluid4 + bafluid shader

[sanyahun]

Hey guys!

Im just started to learn emfluid4 (1 week ago) and now I'm trying to make a match burn. As you can se below in the video, the alpha box is following the animation and it's cut off the flame as reaching it. I don't really know what to change, in the documentation I could'nt find any explanation. Of course if I make any progress I'll upgrade this topic.

The password is: test

Best,

sanyahun

[owei]

Hey..!

Just make the alpha box as big as you need it and chose "Manual, only Alpha Box" as box mode...the automatic box causes the shader to cut off, also the fluid simulation is cutted off. You might also take a look at "Epsilon Cut-Offs (Auto. Box Modes)" if you still want to run the automatic mode... ;)

cheers,
oli

[sanyahun]

Tried it before. When applying the manual, alpha box or all boxes the simulation time multiply by 10 or so I can't change the size of the alphabox only the fluid box. Epsilon cut velocity did the job Thanks

Hey..!

Just make the alpha box as big as you need it and chose "Manual, only Alpha Box" as box mode...the automatic box causes the shader to cut off, also the fluid simulation is cutted off. You might also take a look at "Epsilon Cut-Offs (Auto. Box Modes)" if you still want to run the automatic mode... ;)

cheers,
oli

[owei]

...well, manual means, that the bounding box has to stay the same though out the simulation. You will just define the limits for the hole sim. If you move just the emitter (in the fluid box) and choose a fluid box that is big enough, that would be the workflow ;)

cheers,
oli

[sanyahun]

Just a little update:



Password is the same: test

I hope I'm on the right path

owei: Understand that, the fluid box was big enough, there is a basic ice particle system what is connected to the emfluid (thats the emitter), my guess is that the "visual alpha feedbox" size is the exact same size as the ice pointcloud.

cheers,

sanya

[caledonian_tartan]

...the alpha box is following the animation and it's cut off the flame as reaching it...

i do have the same problem. setting low epsilon cut-offs made it better (and much slower). but still, the particles go through the top of the VisualFeedback_AlphaBox. especially with low cell size.

wasn't there a ice fix once, that we could put on the VisualFeedback_AlphaBox?
should i translate the actual pointcloud's bounding box to SRT?


looks like the auto box is placed at 0,0,0.
with cellsize of 1 it's also too big.
help! Eric!!

manual boxes may work, but it gets a trillion times slower...

any advice?

[owei]

Hey!

Well, these are DIFFERENT problems. One was about voxel (and there shading/simulation cut-offs) and you problem is about particles and their velocities. What do you want to achieve in the end? As the alpha box is a visualization for the actual voxel-grid´s size, it can´t set velocities for the particles outside of it. As the size may change during simulation, it might be that particle won´t get any new velocity for example...so tey will keep on traveling with their current velocity...

cheers,
oli

[caledonian_tartan]

hey oli
i'm sorry i have no clue what you mean. i still think its the same issue. anyhow i think the fact that the autobox isn't behaving right makes it quite similar.

what i meant is to simulate and render a simple emFluid pointcloud:

Get->Primitive->emFluid->Standard with autobox
Set cell size to 0.2
press play and watch how wrong the autobox behaves. particles go through the top. which means they aren't influenced by the velocity grid anymore AND they don't render over there.


now i built this. it helps when used on the autobox insted of standart SRT comming from emFluid.

Get Bounding Box from PointCloud.xsicompound

(13.32 KiB) Downloaded 150 times

EDIT: ok, now i see. this is rubbish, as the voxelgrid does not change, when we fix the autobox SRT here.

[owei]

No idea? That´s bad ;)
Well...do you render particles or voxel?? If you are using BA shaders, I assume you render voxel...So I would recommend switching off the automatic particle emission and just use the show grid functions to see what is happening..

[caledonian_tartan]

i'm using BA (for this example) and also VDB.

i'm rendering voxels and don't have any standard particles in there. only waste


another very strange thing, for which i also blame the autobox is this:
i get completely different simulations with smaller cell size. smaller sizes compress the whole thing at the top.
i thought it should stay the same and just get more detailed...
both are rendered with ba_fluid at frame 77. plus Open GL and a VDB render for reference.

EDIT:
on a second attemt it simulated quite well. still little cut on top. voxelgrid is too small.

[owei]

..well..I´ll try once more.
1. As you use "create particles at flamepositions" in the particle emitter compound you have particles in you scene. They are automatically emitted by emF4 AND advected! So switch them off!
2. go into the emFluid4_Main compound and switch on "visualize grid cells"
3. dig into the emFluid4 Solver compound (open it up with the little "e") and go into the debug compound
4. decrease the velocity value to zero, so that you left with waste at 1.
5. if you now play back, you should get a representation of what is going on in your waste grid
6. adjust the the waste epsilon cut-off in the solver compound.. or use a manual box
7. simulation time will increase, because emFluid has to calculate more stuff because of a lowered cut-off
8. if you decrease the cell size, simulation will change also..it´s a common problem of grid based fluid solvers like emFl4
9. changing simulations due to higher resolutions has nothing (at least a little) to do with the automatic box...what you mean is some kind of detail enhancement which is not supported by emfluid4. Re-simulation is something different..

hope this helps..

[caledonian_tartan]

hi owei

thank you very much for your feedback!
as i believe i'm using one of your examples as a base of my scene, many of this stuff is already tweaked the right way

1. i never had "particle emitter" in my setup. deleted that quite soon... is it better to keep it and switching off?
2. i did quite a few times already, to check the grids behaviour.
3. had it like this
4. had it like this
5. yeah the blue points
6. yes, tried that. helped a bit, but did not completely solve the issue...
7. ok
8. what a pity...
9. ok

[owei]

...no, if there is no emitter, it´s OK...there WERE particles in the image you posted before, right ;)

So if you have everything setup correctly, you should be able to see if your waste is cutted off or not...I guess you try render values that are already cutted off by the alpha box.
So you will have to tweak the display to get correct visualization for that..so set higher values than one in the debug compound for the waste...something like 10 or so..

cheers,
oli

[caledonian_tartan]

...there WERE particles...

this was a little side test, sorry for the confusion.

no succsess with changing the value.
here's a test scene if you like.

after frame 20 it begins to interfere with the top of the bounding box. as the rendering(ba_fluid) does too.

[owei]

I see...so the only chance is to use the the manual alpha box as limit.
If trying with automtic box, maybe also give waste damping value..also lower the follow velocity value in the behavior compound..maybe fiddle around with that. You might also try to lower the velocity cut-off...dig into the solver compound and play with the fluid-box values (value and velocity). Maybe use a higher subframe sampling..(delete simulation stack, add a simed ICE tree again, then you´ll get the property..).

[caledonian_tartan]

hey owei
i'm on the emFluid5 beta now, so let's pretend those posts never happened and see you over there!
thanks anyway